In 1965, Nucor Corporation started making steel in a converted joist plant in Darlington, South Carolina. The product was rebar — the lowest-margin, lowest-quality segment of the steel market. U.S. Steel, Bethlehem Steel, and the other integrated producers didn't care. Rebar was a nuisance product with thin margins and customers who bought on price alone. The integrated mills were happy to cede it.
That indifference was rational. The integrated producers' best customers wanted structural beams, sheet steel, and specialty alloys — high-margin products that justified the enormous fixed costs of blast furnaces and integrated supply chains. Investing capital in rebar production would have meant diverting resources from more profitable segments. Every business school metric — customer satisfaction, gross margin analysis, capital allocation frameworks — confirmed the same conclusion: let the minimills have rebar.
By 1986, Nucor and its fellow minimills controlled 90% of the North American rebar market. Then they moved upmarket to angle iron, bars, and rods. The integrated producers retreated again — those were also low-margin products. By the early 1990s, the minimills had reached structural beams. By the 2000s, they were producing flat-rolled sheet steel, the highest-margin segment the integrated producers had left. Bethlehem Steel filed for bankruptcy in 2001. U.S. Steel survived only through radical restructuring.
This is the pattern Clayton Christensen identified in The Innovator's Dilemma (1997) and named disruptive innovation: a new entrant introduces an inferior product at the bottom of a market, where incumbents' profit margins are thinnest and attention lowest. The product is worse on every dimension the mainstream market cares about. But it's simpler, cheaper, or more accessible — and it serves customers the incumbents are happy to ignore.
Then the disruptive product improves. Year after year, it gets better, creeping upmarket. Each incremental improvement is individually unthreatening — the incumbents still have a superior product for their best customers. But the trajectory is relentless.
At some point the disruptive technology becomes "good enough" for the mainstream market. Not better — just good enough. And at that moment, the price and convenience advantages become decisive. The incumbents discover that the cost structures, business models, and organizational capabilities they built for the premium market are liabilities in the new landscape. Their overhead was calibrated to high-margin products. Their sales teams were trained to sell premium features. Their culture was optimized for precisely the wrong competitive frame.
The critical insight — and the reason Christensen's work endures — is not that incumbents fail because they're poorly managed. They fail because they're well managed. Listening to your best customers, investing in higher-margin products, carefully allocating capital to the most profitable opportunities — these are the behaviors that business schools teach and investors reward. They are also the behaviors that systematically blind organizations to disruptive threats from below.
This is the paradox at the heart of the model: the better a company executes its current strategy, the more vulnerable it becomes to disruption. Excellence and vulnerability are not opposites. They are the same thing, viewed from different time horizons.
Kodak didn't ignore digital photography because its engineers were incompetent. Steve Sasson, a Kodak engineer, built the first digital camera in 1975. Kodak's management made a rational calculation: the company earned $5.6 billion in revenue from film and processing in the mid-1990s, with gross margins above 60%. Digital photography produced images vastly inferior to 35mm film and appealed primarily to early-adopting technologists. Why cannibalize a $5.6 billion business to pursue a low-quality product in a small market? The math was clear. The math was also the trap. Kodak filed for bankruptcy in January 2012.
The pattern is the same whether you look at disk drives, steel, photography, bookselling, or video rental. Barnes & Noble in 1997 had 1,009 superstores and $3.5 billion in annual revenue. Amazon had $148 million and couldn't offer the browsing experience, knowledgeable staff, or instant gratification of a physical bookstore. Barnes & Noble's CEO, Leonard Riggio, told reporters he would "destroy" Amazon. Twenty-five years later, Barnes & Noble operates fewer than 600 stores. Amazon's annual revenue exceeds $570 billion. The incumbent's confidence was proportional to how badly it misread the threat.
Section 2
How to See It
Disruptive innovation operates through a specific sequence: an inferior product enters at the low end, improves faster than the market demands, and eventually displaces the incumbent from below. The signals are predictable once you know what to watch for.
Technology
You're seeing Disruptive Innovation when a new product is dismissed as a toy by industry experts and existing customers. Personal computers in 1977 couldn't match minicomputers on any performance metric. Ken Olsen, founder of Digital Equipment Corporation, reportedly said that year there was "no reason anyone would want a computer in their home." DEC filed for bankruptcy in 1998. The signal isn't the product's current capability — it's the trajectory of improvement relative to what the mainstream market requires.
Business
You're seeing Disruptive Innovation when an incumbent's best customers actively discourage the company from pursuing the new technology. Blockbuster's most profitable customers — families renting new releases on Friday nights — wanted larger stores, more copies of popular titles, and a better in-store experience. None of them were asking for a DVD-by-mail service with no late fees. When Reed Hastings offered to sell Netflix to Blockbuster for $50 million in 2000, Blockbuster's leadership declined. Their customers hadn't asked for it.
Investing
You're seeing Disruptive Innovation when financial analysts dismiss a new entrant because its unit economics don't work at the incumbent's scale. Amazon's book business in 1997 had lower gross margins, higher per-unit logistics costs, and no prospect of matching Barnes & Noble's revenue per square foot. Wall Street analysts pointed to these metrics as proof that online retail was uneconomical. The analysis was technically correct — for the 1997 cost structure. It missed that the cost structure was improving at 40% per year while physical retail costs were fixed.
Markets
You're seeing Disruptive Innovation when a product creates new consumers rather than stealing existing ones. M-Pesa, launched in Kenya in 2007, didn't compete with banks for existing customers. It brought financial services to 17 million Kenyans who had no bank account. By 2023, M-Pesa processed over $314 billion in annual transactions across seven African countries. Traditional banks initially ignored it as a niche telecom product. The pattern is identical to Nucor's rebar: serve the customers nobody wants, then move upmarket.
Section 3
How to Use It
Decision filter
"Is a competitor serving customers we consider unprofitable with a product we consider inferior? Is that product improving faster than our customers' needs are escalating? If both answers are yes, we are being disrupted — regardless of how strong this quarter's earnings look."
As a founder
The most reliable entry strategy in a market with entrenched incumbents is to find the segment they actively don't want. Christensen's data across dozens of industries shows the same pattern: incumbents don't defend the low end because the margins don't justify the effort. Your initial product should be worse than theirs on the metrics their best customers care about — and dramatically better on a dimension those customers dismiss as irrelevant: price, simplicity, accessibility, or convenience.
Netflix's initial DVD-by-mail service had worse selection than Blockbuster for new releases. It was dramatically better on convenience and cost for back-catalog browsing. The key discipline: resist the temptation to compete on the incumbent's terms prematurely. Your goal in year one is to dominate the segment they don't want, not to chase the customers they do. The upmarket migration comes later — once your product has improved enough and your cost structure has hardened into a durable advantage.
As an investor
The earliest financial signal of disruption is not market share loss — it's margin improvement by the disruptor combined with market share stability at the incumbent. The incumbent's numbers look fine right up until the inflection. Blockbuster's same-store revenue was still growing in 2003 when Netflix crossed one million subscribers. Nokia reported record smartphone profits in Q4 2007, the quarter the iPhone launched.
The investor's edge is watching the disruptor's improvement trajectory, not the incumbent's quarterly results. Plot the disruptor's key performance metrics on a logarithmic scale. If the improvement is linear on that scale — consistent percentage gains each year — the technology is on an exponential trajectory that will eventually cross the market's "good enough" threshold. Digital camera resolution improved at roughly 40% per year through the 2000s. The moment that trajectory crossed the threshold for acceptable print quality, Kodak's film business had a precisely datable expiration.
As a decision-maker
If you run an incumbent organization, the standard advice — set up an autonomous unit to pursue the disruptive technology — is correct but insufficient. The autonomous unit needs its own P&L, its own incentive structure, and its own definition of success that doesn't reference the parent company's margins.
IBM's PC division succeeded in the 1980s because it operated independently in Boca Raton, Florida, far from the mainframe organization's gravitational pull. The team sourced components externally, used an open architecture (heresy for IBM), and shipped on a timeline the parent company's process would have made impossible. When IBM reintegrated the PC division into the corporate structure, the mainframe culture slowly strangled it. The lesson: autonomy isn't a nice-to-have. It's the structural precondition for a large organization to pursue a disruptive opportunity without the immune system of the core business killing it.
Common misapplication: The word "disruption" has been drained of precision. Uber is frequently called disruptive, but by Christensen's own analysis, it was not — Uber didn't start with an inferior product in a low-end market. It launched with a superior experience (cleaner cars, GPS tracking, cashless payment) in San Francisco's premium market. That's sustaining innovation: a better product for existing customers. Tesla faces the same misclassification. The Model S debuted as a $70,000 luxury vehicle competing directly with BMW and Mercedes on performance. That's sustaining. Genuine disruption starts below, not above.
The distinction matters because the strategic response to a sustaining innovator is fundamentally different from the response to a disruptive innovator. Against a sustaining innovator, you compete on features and execution — build a better product, move faster, outspend. Against a disruptive innovator, feature competition is irrelevant because the disruptor isn't competing on your features. The response requires structural change: either cannibalize your own business through an autonomous unit, or accept that the low-end segment you're ceding today is the mainstream market of tomorrow.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
Disruption is not a theory about technology. It's a theory about rational decision-making and its blind spots. The founders below didn't succeed because incumbents were incompetent — they succeeded because incumbents were competently optimizing for the wrong competitive frame.
The pattern spans sixty years and five industries — semiconductors, video entertainment, book retail, mobile computing, and artificial intelligence infrastructure. In each case, the entrant started with a product that incumbents correctly identified as inferior, served customers incumbents correctly identified as unprofitable, and operated at margins incumbents correctly identified as unacceptable.
The word "correctly" is doing all the work in those sentences. The incumbents were not wrong about the present. They were wrong about the trajectory — and the organizational structures that made them right about today made them blind to tomorrow.
Grove was one of Christensen's earliest and most important converts. After reading a pre-publication manuscript of The Innovator's Dilemma in 1996, Grove invited Christensen to present to Intel's senior leadership. The problem was AMD's K6 processor — a cheaper, lower-performance chip eating into Intel's share of the sub-$1,000 PC market. Intel's instinct, consistent with disruption theory, was to retreat upmarket and focus on the high-margin Pentium II for premium buyers.
Grove recognized the pattern and reversed Intel's course. He authorized the Celeron processor line in April 1998 — a deliberately stripped-down chip designed to compete with AMD at the low end. Intel's sales force resisted. The margins on Celeron were roughly half those of the Pentium II. Selling a cheaper chip meant cannibalizing Intel's own premium product.
Grove overrode the objections. Within a year, Celeron captured significant share in the sub-$1,000 PC segment and blocked AMD's upmarket advance. By 2000, Intel's total unit shipments had increased, and the Celeron line's volume pushed manufacturing costs lower across the entire product portfolio. Grove had used Christensen's framework to do what disruption theory says incumbents almost never do: cannibalize yourself before a competitor does it for you. The Celeron decision remains the single best documented case of a sitting CEO reading disruption theory and successfully applying it in real time.
Netflix's disruption of Blockbuster unfolded in two distinct phases, each following Christensen's model with textbook precision.
Phase one: DVD-by-mail, launched in 1998. The product was objectively inferior to Blockbuster for impulse rentals — you couldn't walk in Friday night and walk out with a movie. But it was superior on two dimensions Blockbuster's core customers didn't value: no late fees and access to a deep back catalog. Netflix's early subscribers were film enthusiasts and serial renters — customers Blockbuster considered low-value because they held titles longer and generated less late-fee revenue. By 2005, Netflix had 4.2 million subscribers. Blockbuster's revenue was still growing.
Phase two: streaming, launched in 2007. The initial library was small — roughly 1,000 titles versus Blockbuster's 6,500 in-store. Video quality was poor, limited by bandwidth. But the trajectory was steep. Bandwidth costs dropped roughly 30% per year. Netflix's content library expanded as licensing deals multiplied. By 2010, streaming had surpassed DVD-by-mail in engagement. Blockbuster filed for bankruptcy in September 2010 with $900 million in debt.
Hastings understood from the outset that Netflix's DVD business would eventually be disrupted by streaming — and that if Netflix didn't do it, someone else would. He allocated increasing resources to streaming infrastructure even as the DVD business remained profitable, deliberately cannibalizing his own revenue stream. That willingness to self-disrupt is the rarest behavior in corporate strategy.
Amazon's disruption of physical bookstores follows the pattern precisely — with one variation that makes it particularly instructive.
In 1995, Amazon couldn't match Barnes & Noble on any traditional retail metric. No browsing experience. No knowledgeable staff. Delivery delays of days or weeks. Barnes & Noble operated 1,009 superstores with $3.5 billion in annual revenue. Amazon's first-year revenue was $511,000.
But Amazon served a customer segment that Barnes & Noble structurally couldn't: buyers searching for specific titles that no physical store would stock. A Barnes & Noble superstore carried roughly 175,000 titles. Amazon listed 1.1 million. The long tail of obscure, niche, and out-of-print books was unprofitable for physical retail — inventory carrying cost per title exceeded expected revenue. For Amazon, adding a title to its database cost almost nothing. The segment Barnes & Noble rationally ignored was Amazon's entire early market.
The instructive variation: Bezos didn't wait for the book business to mature before disrupting adjacent categories. By 1998, Amazon had expanded into music, DVDs, and electronics — each time entering at the low end of customer experience while being dramatically better on selection and price. Bezos's 1997 letter to shareholders described the strategy explicitly: use the cost structure advantages of online retail to enter one category after another, applying the same disruptive logic that worked in books.
The iPhone's disruption is the most commercially consequential case of the twenty-first century — and the most commonly misunderstood.
The standard narrative says Apple built a better phone and outcompeted Nokia on features. That's sustaining innovation. The disruption happened one layer down: the iPhone disrupted the mobile phone industry structure by transforming the phone from a hardware product into a software platform.
Nokia in 2007 controlled 49.4% of the global smartphone market. The company spent $8.6 billion on R&D that year — more than Apple's entire operating budget. Nokia's phones were reliable, well-engineered, and optimized for the job the mobile industry defined as "a phone": voice calls, text messages, basic data services. The carriers — Vodafone, AT&T, T-Mobile — controlled the customer relationship, the software experience, and the revenue from data services.
Jobs bypassed the entire structure. The App Store, launched in July 2008, created a platform where third-party developers could build software directly for consumers — cutting the carriers out of the software value chain. Nokia's Symbian OS could theoretically support third-party apps, but the carrier-controlled distribution model made development impractical. Apple created a new market for mobile software that Nokia's organizational structure was incapable of addressing. By 2013, Nokia's smartphone market share had fallen below 3%. Microsoft acquired Nokia's phone division for $7.2 billion in September 2013 — a fraction of the $150 billion in market capitalization Nokia had commanded six years earlier.
NVIDIA's trajectory from gaming peripheral to the infrastructure backbone of artificial intelligence is a disruption story the incumbents never saw coming — because the disruptive product wasn't inferior. It was irrelevant.
When Huang co-founded NVIDIA in 1993, the company made graphics processing units for video games. Intel dominated computing. The GPU was a specialized co-processor that Intel's engineers dismissed as a niche accessory — useful for rendering polygons, useless for serious computation. Intel's x86 architecture was the platform that mattered. GPUs sat on the periphery.
The disruption began not in gaming but in scientific computing. Researchers at Stanford and other universities discovered in the mid-2000s that NVIDIA's CUDA programming platform — released in 2006 — allowed GPUs to perform massively parallel calculations thousands of times faster than CPUs for certain workloads. Machine learning, molecular dynamics, fluid simulation — problems that took days on Intel hardware took hours on NVIDIA's. The customer base was tiny: a few hundred academic labs. The revenue was negligible. Intel had no reason to respond.
Then deep learning arrived. Alex Krizhevsky's AlexNet won the ImageNet competition in 2012 using two NVIDIA GTX 580 GPUs — consumer gaming cards costing $500 each. The entire modern AI revolution was seeded on hardware Intel considered a toy. By 2024, NVIDIA's data center revenue reached $47.5 billion in a single fiscal year. Intel's data center group generated $15.5 billion. The gaming co-processor had become the computing platform, and Intel's CPU-centric architecture — the most successful scale position in technology history — was structurally unable to compete on the workloads that now defined the frontier.
Section 6
Visual Explanation
The Disruptive Innovation pattern — entrants start below market requirements, improve faster than customer needs escalate, and eventually overtake incumbents
Section 7
Connected Models
Disruptive innovation doesn't operate in isolation. It interacts with adjacent strategic frameworks in ways that either amplify the disruption pattern or create friction against it. The strongest disruptions in business history — Netflix, Amazon, Nucor — combined the disruptive entry strategy with reinforcing dynamics from other models. The incumbents who fell fastest were those whose existing strategic frameworks actively concealed the threat.
Understanding these connections separates founders who deploy the concept strategically from those who use it as a buzzword.
Reinforces
[Jobs to Be Done](/mental-models/jobs-to-be-done)
Christensen developed Jobs to Be Done as a companion framework — and the two reinforce each other powerfully. Jobs to Be Done explains why customers switch to a disruptive product despite its inferior performance: the disruptive product does a different job, or does the same job in a context the incumbent doesn't serve. Netflix's early subscribers weren't hiring Netflix for Blockbuster's job (Friday night impulse rental). They were hiring it for a different job: curated discovery of films they wouldn't find at Blockbuster. When the disruptor eventually does the incumbent's job well enough, the switch accelerates. Jobs to Be Done predicts which disruptions will cross the chasm; disruption theory predicts the trajectory once they do.
Reinforces
S-Curve
Technology adoption follows an S-shaped curve — slow initial uptake, rapid growth, then saturation. Disruptive innovation theory explains why the S-curve of the new technology eventually crosses the old one. The incumbent's technology follows its own S-curve, approaching diminishing returns at the top. The disruptive technology starts a new curve from a lower base with a steeper slope.
Digital camera resolution in the early 2000s was climbing steeply while 35mm film had long since plateaued. Streaming video quality improved along a predictable S-curve while DVD quality was fixed by the physical medium. The crossing point — when the new S-curve surpasses the old — is the moment of disruption. Christensen's performance trajectory charts are, mathematically, overlapping S-curves viewed from the demand side. The incumbent sees a flat line. The disruptor sees a ramp. Both are right about the present. Only one is right about the future.
Tension
[Economies of Scale](/mental-models/economies-of-scale)
Section 8
One Key Quote
"The logical, competent decisions of management that are critical to the success of their companies are also the reasons why they lose their positions of leadership."
— Clayton Christensen, The Innovator's Dilemma (1997)
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Disruptive innovation is simultaneously the most cited and most misapplied framework in business strategy. Half of Silicon Valley pitch decks claim to be "disrupting" something. Roughly 5% describe actual disruption by Christensen's definition. The rest are building better products for existing customers — sustaining innovation with disruptive branding. The misuse has become so pervasive that Christensen himself spent the last decade of his career trying to reclaim the term's precision, publishing a 2015 Harvard Business Review article titled "What Is Disruptive Innovation?" specifically to correct the dilution.
The distinction is not semantic. It's strategic. If you're building a sustaining innovation, your competition is primarily on execution — features, speed, marketing, distribution. If you're building a genuinely disruptive innovation, your advantage is structural: incumbents will rationally choose not to compete with you because the economics of your market don't justify their cost structure. That structural advantage is what makes true disruption defensible. A better product can be copied. A business model that incumbents can't adopt without destroying their existing margins cannot.
The pattern I watch for most carefully: the incumbent's emotional response. When a large company dismisses a new entrant with contempt — "their product is terrible," "their customers are low-value," "they'll never serve our market" — those are the precise signals Christensen's theory predicts. Dismissal is the rational response that precedes disruption. Kodak's internal assessment of digital photography in the mid-1990s used all of those phrases. So did Blockbuster's assessment of Netflix. So did Nokia's assessment of the iPhone.
What gets lost in popular accounts: the timescale. Christensen's steel minimill case unfolded over thirty years. Netflix took thirteen years from founding (1997) to Blockbuster's bankruptcy (2010). Amazon took over two decades to displace physical retail as the default shopping channel for most American households. Disruption is not a sudden event. It's a slow-motion displacement that appears sudden only because the incumbent's collapse happens rapidly once the disruption point is reached — the long fuse, the fast explosion.
The hardest practical question is self-disruption. Christensen's framework implies that the rational response for an incumbent is to build an autonomous unit that cannibalizes its own business before a competitor does. Intel did this with Celeron under Grove. Netflix did this with streaming under Hastings. Apple did this when the iPad cannibalized Mac sales. These are extreme outliers. The organizational antibodies against self-cannibalization are powerful: sales teams that lose commissions, divisions that lose budgets, executives that lose headcount. Most incumbents can't override those incentives even when they understand the theory intellectually.
Section 10
Test Yourself
Disruption is overdiagnosed. Every startup claims it. Every incumbent fears it. Most of the time, neither is right.
The scenarios below test whether you can distinguish genuine disruptive innovation — the specific pattern Christensen identified — from sustaining innovation, competitive displacement, and ordinary market competition wearing the "disruption" label. The most common error: confusing "new and successful" with "disruptive." Google, Uber, and Tesla are all extraordinary companies. None launched as disruptive innovators by Christensen's definition. Precision matters — the framework's predictive power depends on applying it to the right situations.
Is this mental model at work here?
Scenario 1
A startup launches an electric vehicle priced at $110,000, targeting luxury buyers with 0–60 mph acceleration faster than any existing sports car. The vehicle wins design awards and is reviewed favorably by automotive journalists who compare it to Porsche and BMW.
Scenario 2
A steel company builds small electric arc furnaces that melt scrap metal into rebar. The product quality is low — inconsistent gauge, surface imperfections, unsuitable for structural applications. Integrated steel producers happily cede the rebar market because margins are 7% versus 25% on structural steel. Over fifteen years, the company improves its process and begins producing structural beams, then sheet steel.
Scenario 3
A search engine delivers dramatically better results than the existing market leader, Yahoo. The product serves the same users searching for the same information — but does it faster, with cleaner design and a more effective ranking algorithm. Within five years, the new entrant dominates the market.
Scenario 4
A mobile payment service launches in an emerging market where 80% of the population has no bank account. The service allows users to send money via basic SMS on feature phones. Transaction limits are low, the interface is primitive, and the service offers none of the features — credit, savings, investments — that banks provide. Banks dismiss it as a telecom gimmick. Within six years, it processes more transactions than all the country's banks combined.
Section 11
Top Resources
The literature on disruption splits cleanly into two categories: Christensen's own work and rigorous refinements, versus the popular dilutions that stripped the theory of its precision.
Start with The Innovator's Dilemma for the diagnosis and The Innovator's Solution for the prescription. Grove's memoir is the essential practitioner's companion — the rare account of someone who read the theory and applied it under pressure. The 2015 HBR article is required corrective reading for anyone who's been using "disruptive" loosely. Ignore anything that uses the word as a synonym for "new."
The foundational text. Christensen's disk drive industry data is the most granular empirical evidence for the theory, and the steel minimill case study remains the clearest illustration of the mechanism. The book's power comes from its central paradox: good management causes failure. Read the original rather than summaries — the nuances that popular accounts strip away are the nuances that make the framework useful.
The prescriptive companion to the Dilemma. Where the first book diagnosed the problem, this one proposes solutions: how incumbents can create disruptive growth businesses, how to identify which customers will adopt first, and how to structure autonomous units. The chapter on Jobs to Be Done is where Christensen first introduced the framework that became its own field of practice.
Grove's account of navigating Intel through "strategic inflection points" — moments when a business's underlying assumptions change fundamentally. Written the year before The Innovator's Dilemma, Grove's framework is complementary: he describes how incumbents can recognize and respond to disruption. The Celeron case, decided after Grove read Christensen's manuscript, makes this the rare strategic text written by someone who both understood the theory and applied it under real pressure.
Christensen's own correction to twenty years of misapplication. Published in Harvard Business Review, the article explicitly addresses what disruption is not — including the argument that Uber is not disruptive by the theory's definition. Essential for anyone who wants to use the framework with precision rather than as a marketing label.
The full articulation of Jobs to Be Done theory — the framework Christensen developed to explain why customers adopt disruptive products. The milkshake case study (why do people buy milkshakes at 6:30 AM?) demonstrates how understanding the job a product does — rather than the customer's demographics — predicts adoption. Pairs directly with disruption theory to explain both the demand-side and supply-side dynamics of displacement.
Scale economics should protect incumbents — and often do against sustaining innovators. But disruption theory reveals how scale advantages become vulnerabilities when the production technology shifts. U.S. Steel's massive blast furnaces and integrated supply chain were optimized for a specific way of making steel. When Nucor's electric arc furnaces changed the cost function entirely, U.S. Steel's scale became an anchor — not a shield.
The capital invested in blast furnace infrastructure couldn't be redirected to minimill technology without writing off billions in sunk costs. Kodak's film manufacturing scale — the largest in the world — became worthless when digital photography eliminated demand for film. Blockbuster's network of 9,000 retail locations became a $1 billion liability when streaming eliminated the need for physical stores. Scale advantages are durable only as long as the underlying production technology remains constant. Disruption changes the technology, resetting the scale equation entirely.
Tension
Porter's Five Forces
Michael Porter's framework analyzes industry structure — buyer power, supplier power, threat of substitutes, barriers to entry, rivalry among competitors. Disruption reveals a blind spot: Porter's analysis assesses threats within the existing industry definition. When Christensen asked Porter why Five Forces didn't predict integrated steel's disruption by minimills, the answer was that minimills were in a "different industry" making a different product. By the time they were clearly in the same industry, the disruption was complete.
The tension is methodological: Five Forces excels at analyzing competition within a stable industry structure. Disruption changes the structure itself. A Five Forces analysis of the video rental industry in 2002 would have rated Blockbuster's competitive position as strong — high barriers to entry, significant supplier leverage, fragmented rivalry. It would have missed Netflix entirely because Netflix, at that point, wasn't in the video rental industry as Porter's framework defined it.
Leads-to
First Principles Thinking
Disruption frequently originates from someone applying first principles reasoning to an industry's cost structure or value chain. Nucor's Ken Iverson asked: what is the minimum infrastructure required to produce steel? An electric arc furnace, scrap metal, and a small crew. Everything else — blast furnaces, coke ovens, ore mines, railroad fleets — was the integrated model's convention, not physical necessity.
Reed Hastings asked the same kind of question about video distribution: what does it actually cost to deliver bits over wire? Jeff Bezos asked it about book retail: what does it cost to list one more title in a database? The disruption followed from the decomposition in each case. First principles identifies the opportunity — the gap between what something costs by convention and what it costs by physics. Disruption theory predicts the competitive dynamics that unfold once someone acts on that gap.
Leads-to
[Network Effects](/mental-models/network-effects)
Many successful disruptors leverage network effects to accelerate their upmarket march once they cross the performance threshold. The iPhone's App Store is the clearest example: once the platform attracted enough users to interest developers, and enough developers to interest users, the resulting network effects created an advantage that Nokia's superior hardware engineering couldn't overcome. By 2010, the App Store had over 225,000 apps — a software ecosystem no competitor could replicate through engineering effort alone.
Amazon Marketplace followed the same pattern — third-party sellers attracted buyers, buyers attracted sellers, and the cycle produced a selection advantage no physical retailer could match regardless of store count or purchasing power. Disruption creates the opening. Network effects close the door behind the disruptor, converting an initial cost-structure advantage into a demand-side moat that compounds with each additional user.
One pattern underweighted in Christensen's original work: disruption through business model, not technology. Southwest Airlines didn't have better planes than American or United. It had a different business model — point-to-point routing, single aircraft type, no assigned seats, no hub-and-spoke complexity. The result was a cost per available seat mile roughly 30% below legacy carriers. The legacy carriers couldn't match it without dismantling their hub operations, frequent flyer programs, and first-class cabins — the core of their most profitable customer relationships. Business model disruption is often more durable than technological disruption because the barriers to imitation are organizational, not technical.
The most underappreciated dimension: speed of disruption is accelerating. Christensen's steel minimill case took thirty years from Nucor's rebar entry to Bethlehem Steel's bankruptcy. Netflix took thirteen years from founding to Blockbuster's collapse. The iPhone took six years from launch to Nokia's acquisition by Microsoft. AI coding tools moved from novelty to production adoption in roughly two years. Each successive wave of disruption compresses the timeline because the rate of technology improvement is itself accelerating. The practical implication: incumbents have less time to recognize and respond to disruptive threats than Christensen's historical case studies suggest. The thirty-year warning window of the steel era is now closer to five.
The current frontier of disruption is AI's impact on knowledge work. The pattern is visible and accelerating. GitHub Copilot in 2022 couldn't write production-quality code — it was a toy that generated plausible-looking snippets requiring heavy human editing. By 2024, AI coding tools were handling 30-40% of code at companies like Google and Meta. The trajectory matches Christensen's model exactly: an initially inferior product improving faster than the market's threshold of "good enough." The incumbents being disrupted aren't software companies — they're the traditional structures of professional services, legal research, financial analysis, and content production. The firms that dismiss AI output as "not good enough for our clients" are reciting Blockbuster's lines with different nouns.
My honest read: disruption theory is one of the three or four indispensable frameworks for understanding competitive dynamics. Apply it only to situations where a new entrant is entering from below with an inferior product that's improving faster than market demands escalate. Apply it honestly — including to your own company, where the temptation is always to believe you're the disruptor rather than the disrupted. And apply it with patience. The founders who successfully execute disruptive strategies are the ones willing to be dismissed as irrelevant for years before the market catches up to their trajectory.
